> ## Documentation Index
> Fetch the complete documentation index at: https://docs.snowdoughnut.com/llms.txt
> Use this file to discover all available pages before exploring further.

# What this framework covers

> The line between what this knowledge base standardizes and what stays inside your ad platforms and with your media buyer. Read it before your first campaign so you can tell a deliberate boundary from a missing page.

This knowledge base is an operations framework. It fixes the decisions that have to be reviewable: what a thing is called, what it is optimizing toward, what it cost, who approved it, and where the record was filed. Work through it once and every campaign after that has something to check against.

It is not platform training. Nothing here teaches you to research keywords, write a headline, install a conversion tag, or configure a Google Ads account. Those are real skills, they are documented by the platforms themselves, and they change faster than any internal knowledge base can track.

<Note>
  The gaps below are deliberate. If a topic is listed here as out of scope, there is no missing page to go and find.
</Note>

## The split

| This knowledge base                                | Your ad platform and your media buyer               |
| :------------------------------------------------- | :-------------------------------------------------- |
| What the campaign is for, and why this type        | Which keywords, bids, and placements deliver it     |
| Which conversion action counts as a result         | How that conversion action gets installed and fires |
| The name every object carries, at all three levels | The buttons that create the object                  |
| What was approved, by whom, on what date           | What gets adjusted between reviews                  |
| Where the finished record is filed                 | Where the live data sits                            |

The pattern repeats: this knowledge base records the **decision** and the **name**. The platform holds the **execution**.

## Where the boundary falls in practice

Each row below is a topic our briefs and checklists refer to without explaining. The middle column is all this framework asks of you. The right column is who makes the call.

| Topic                                              | What a brief or checklist asks for                                              | Who owns it                                |
| :------------------------------------------------- | :------------------------------------------------------------------------------ | :----------------------------------------- |
| Keyword research and selection                     | Nothing. Ad groups are briefed as themes, not keyword lists                     | Media buyer                                |
| Match type policy                                  | One line recording the policy you chose and why                                 | Media buyer, recorded by the owner         |
| Negative keyword lists                             | Confirmation that your standing list is applied                                 | Media buyer, list lives in the ad platform |
| Ad copy craft                                      | A [copy brief](/copy-brief) and a filled copy template within character limits  | Copywriter                                 |
| Conversion tracking setup                          | Confirmation the action imports and fires, and how that was verified            | Whoever owns your analytics stack          |
| Consent mode, enhanced conversions, offline import | Nothing. Note in the brief if the campaign depends on one                       | Analytics or web owner                     |
| Account structure, access, billing, MCC            | Nothing. The brief assumes a working account exists                             | Account owner                              |
| Shared budgets and portfolio bid strategies        | Nothing, unless one changes what this campaign can spend                        | Media buyer                                |
| Auto-applied recommendations                       | A recorded decision on which are on or off, because they rewrite briefed values | Media buyer, recorded in the brief         |
| Bid and target adjustments after launch            | Nothing between reviews. The starting value is briefed                          | Media buyer                                |
| Asset rotation                                     | Nothing. Approved assets come from the creative brief                           | Media buyer                                |

<Warning>
  Auto-applied recommendations are the one out-of-scope item that can quietly undo in-scope work. Left on, they can rewrite ad copy, add keywords, and change bid strategies without anyone approving the change, which breaks the rule that a briefed value is what is actually running. Deciding they stay on is fine. Not knowing is not.
</Warning>

## Why the boundary sits here

Two reasons, and both are practical rather than principled.

**Platform mechanics go stale.** A page explaining how to install a conversion tag is wrong within a year, and a wrong internal page is worse than no page, because people trust it. Every platform detail this knowledge base does carry, campaign types, bid strategy names, creative specs, is dated on the page and reviewed against the platform.

**Judgement does not compress into a checklist.** Which keywords to add, when to move off a maximize strategy, whether a weak ad strength is acceptable: those depend on live data that only the person watching the account can see. A framework that pretended to answer them would be overridden in week two and ignored from then on.

What survives is the layer above: the decision was made deliberately, it was written down before spend started, and the result can be found later.

## What to do when you hit the edge

<Steps>
  <Step title="Check whether a decision is being made">
    If the answer changes what the campaign is optimizing toward, what it spends, or what it is called, it belongs in the brief. If it changes how a briefed value gets delivered, it does not.
  </Step>

  <Step title="Record the decision, not the click path">
    "Exact and phrase only, broad off until CPA is stable" is a decision. "Open the keyword planner and click Add" is a click path, and it belongs in the platform's own documentation.
  </Step>

  <Step title="Add a best practice if the answer keeps recurring">
    A choice your team keeps making the same way is a candidate for a [best practice](/campaign-structure-best-practices) page, which records the default without pretending to be a tutorial.
  </Step>

  <Step title="Raise it if the boundary is wrong">
    The split above is a decision, not a law. If a gap is costing your team real time every campaign, it can move.
  </Step>
</Steps>

## Related resources

* [**Get started**](/overview) The seven setup stages, in order.
* [**Setup checklist**](/setup-checklist) Track your progress through them.
* [**Make a campaign**](/make-a-campaign) The most common task once setup is done.
* [**Campaign structure best practices**](/campaign-structure-best-practices) The defaults that sit just inside the line.
* [**Glossary**](/glossary) Every term used across these documents.